Share Email Print
cover

Proceedings Paper

Subpixel boundary backward substitution reconstruction algorithm for not uniform microscan to FPA and blind micromotion matching
Author(s): Yi-nan Chen; Wei-qi Jin; Lei Zhao; Mei-jing Gao; Lin Zhao
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

For the subpixel micro-scanning imaging, we propose the reconstruction algorithm based on neither interpolation nor super-resolution idea but one of the block-by-block method recursive from the boundary to centre when additional narrowband boundary view-field diaphragm whose radiation is known a prior. The aim of the predicted boundary value is to add the conditions for solving the ununiqueness ill-problem to the inverse transition matrix from the destructed process. For the non-uniform scan factor, the improved algorithm associated with certain non-uniform motion variables is proposed. Additionally, attention is focused on the case of unknown subpixel motion, when the reconstructed images are blurred by motion parameter modulation and neighbouring point aliasing because the value of micro-motion is not the correct one. Unlike other methods that the image registration is accomplished before multi-frame restoration from undersampled sequences frame by frame, in this paper, 2-D motion vector is estimated in single frame just from the blur character of reconstructed grids. We demonstrate that once the estimated motion approaches to the real one, square summation of all pixels over the unmatched image approximately descends to the minimum. The matching track based on recursive Newton secant approaching is optimized for high matching speed and precision by different strategies, including matching region hunting, matching direction choosing and convergence prejudgement. All iterative step lengths with respect to motion parameters are substituted by the suitable values derived from the statistic process and one or multi-secant solution. The simulations demonstrate the feasibility of the matching algorithm and the obvious resolution enhancement compared to the direct oversampling image.

Paper Details

Date Published: 8 January 2008
PDF: 12 pages
Proc. SPIE 6835, Infrared Materials, Devices, and Applications, 68350O (8 January 2008); doi: 10.1117/12.754499
Show Author Affiliations
Yi-nan Chen, Beijing Institute of Technology (China)
Wei-qi Jin, Beijing Institute of Technology (China)
Lei Zhao, Beijing Institute of Technology (China)
Mei-jing Gao, Beijing Institute of Technology (China)
Lin Zhao, Beijing Institute of Technology (China)


Published in SPIE Proceedings Vol. 6835:
Infrared Materials, Devices, and Applications
Yi Cai; Haimei Gong; Jean-Pierre Chatard, Editor(s)

© SPIE. Terms of Use
Back to Top